Processor and method of controlling execution of processes

This application is a continuation application of U.S. application Ser. No. 13/631,272 filed Sep. 28, 2012, which is a continuation of application Ser. No. 11/474,948 filed Jun. 27, 2006, which claims priority from Japanese Patent Application No. 2005-187641 filed on Jun. 28, 2005 in the Japanese Patent Office, the disclosures of which are incorporated herein by reference in their entirety.

1. Field of the Invention

The present invention relates to a processing technique when a fault has occurred in a processor for executing a plurality of processes at a same time.

2. Description of the Related Art

A processor with a single CPU had been generally used for a control process. However, since the control process becomes higher in level and more complex, the single CPU has become impossible to execute the whole of process. For this reason, a multi-processor system is developed where a plurality of CPUs execute a plurality of processes at a same time. Here, it is supposed that the multi-processor system executes a process A of a process A1 and a process A2. Also, it is supposed that the multi-processor system has a CPU 1 and a CPU 2 and they execute the process A1 and the process A2, respectively. In this case, if the CPU 1 and the CPU 2 can normally execute the processes A1 and A2, the process A is completed without any problem. However, in such a multi-processor system, if one of the plurality of CPUs is inoperable, the entire system is down. For example, when the CPU 1 is in a frozen state or in a fault state in which an indefinite loop is executed, the process A1 to be executed by the CPU 1 is not completed. For this reason, even if the CPU 2 can normally complete the process A2, the process A is not completed. Also, since the process A is not completed, the CPU 2 cannot start a next process even if completing the process A2. In this way, the entire multi-processor processor cannot be normally operated.

The technique related to a multi-processor system is disclosed in Japanese Laid Open Patent Publication (JP-P2000-76199A). In this conventional example, a CPU issues a request signal to an arbitrator circuit. The arbitrator circuit executes a bus arbitration and gives a bus use right (grant signal) to the CPU. When the bus is released, the CPU issues a completion signal to the arbitrator circuit. Then, the bus is released in accordance therewith. In this way, the switching between the CPUs (the control of the occupation/relief of input/output I/F) is executed by reserving the bus use right.

It is therefore an object of the present invention to provide a processor where, even if one of a plurality of CPUs is inoperable, the process of the CPU can be executed by a different CPU.

In an aspect of the present invention, a processor includes a plurality of processing sections, each of which executes a predetermined process; a plurality of fault detecting circuits respectively provided for the plurality of processing sections, to detect a fault in one of the plurality of processing sections as a fault processing section to generate a fault detection signal; and a fault monitoring and control section configured to control a normal processing section as at least one of the plurality of processing sections other than the fault processing section to execute a relieving process in response to the fault detection signal.

Here, the relieving process may be determined based on a process load of the fault processing section, a process load of the normal processing section, and priority levels of processes to be executed by the fault processing section and the normal processing section.

In this case, the normal processing section may execute the process to be executed by the fault processing section in the relieving process. Also, the normal processing section may execute one of the process to be executed by the fault processing section and the process to be executed by the normal processing section in the relieving process based on the priority levels.

Also, the processor may further include a single input and output terminal; and a selector section connected with the plurality of processing sections and the input and output terminal and configured to connect the input and output terminal with the normal processing section in response to a selection control signal. The fault monitoring and control section may generate the selection control signal in response to the fault detection signal.

Also, each of the plurality of processing sections may include a CPU; a bus connected with the CPU; and at least one input/output interface connected with the bus. Each of the plurality of fault detecting circuits may include a first fault detecting circuit provided for the CPU; a second fault detecting circuit provided for the bus; and a third fault detecting circuit provided for the input/output interface.

Also, the fault monitoring and control section may include a detection signal decoder configured to decode the fault detection signal to generate a decoding resultant signal; and a storage section configured to store the decoding resultant signal. In this case, the processor may further include a terminal used to output the decoding resultant signal. Also, the fault monitoring and control section may further include an access section configured to access the storage section in response to a request from the normal processing section and to transmit a data of the fault processing section to the normal processing section.

Also, the processor may further include a register provided for each of the plurality of processing sections to hold a reset instruction and configured to output the set reset instruction to the fault processing section when the fault has occurred in a corresponding processing section.

Also, the processor may further include a plurality of clock signal generators respectively provided for the plurality of processing sections to generate a plurality of clock signals for the plurality of processing sections; and a voltage detecting circuit configured to detect that a power supply voltage is lower than a threshold value and to output the fault detection signal to the fault monitoring and control section. The fault monitoring and control section regards the processing section which operates in one of the plurality of clock signals which has a higher frequency, as the fault processing section, and the processing section which operates in one of the plurality of clock signals which has a lower frequency, as the normal processing section, when the detected power supply voltage is lower than the threshold value, and controls the normal processing section to execute the relieving process.

Also, in another aspect of the present invention, a method of controlling execution of processes by a processor comprising a plurality of processing sections, is achieved by detecting a fault in each of the plurality of processing sections as a fault processing section to generate a fault detection signal; and by controlling a normal processing section as at least one of the plurality of processing sections other than the fault processing section to execute a relieving process in response to the fault detection signal.

Here, the method may be achieved by further determining the relieving process based on a process load of the fault processing section, a process load of the normal processing section, and priority levels of processes to be executed by the fault processing section and the normal processing section.

Also, the controlling may be achieved by controlling the normal processing section to execute a process to be executed by the fault processing section in the relieving process.

Also, the controlling may be achieved by controlling the normal processing section to the normal processing section to execute one of a process to be executed by the fault processing section and a process to be executed by the normal processing section in the relieving process based on priority levels of the processes.

Also, the method may be achieved by further generating a selection control signal in response to the fault detection signal; and connecting an input and output terminal with the normal processing section in response to the selection control signal.

Also, the method may be achieved by further decoding the fault detection signal to generate a decoding resultant signal; storing the decoding resultant signal in a storage section; and outputting the decoding resultant signal from the storage section.

Also, the method may be achieved by further accessing the storage section in response to a request from the normal processing section; and transmitting a data of the fault processing section to the normal processing section.

Also, the method may be achieved by further setting a reset instruction in registers; and outputting the reset instruction to the fault processing section when the fault has occurred in a corresponding processing section.

Also, the detecting may be achieved by generating a plurality of clock signals for the plurality of processing sections; and by detecting that a power supply voltage is lower than a threshold value, to output the fault detection signal. The controlling may be achieved by regarding the processing section which operates in one of the plurality of clock signals which has a higher frequency, as the fault processing section, and the processing section which operates in one of the plurality of clock signals which has a lower frequency, as the normal processing section, when the detected power supply voltage is lower than the threshold value; and by controlling the normal processing section to execute the relieving process.

FIG. 1 is a block diagram showing a configuration of a processor according to a first embodiment of the present invention;

FIGS. 2A to 2C are timing charts showing an operation of the processor according to the first embodiment of the present invention;

FIG. 3 is a flow chart showing an operation of the processor according to the first embodiment of the present invention;

FIG. 4 is a block diagram showing the configuration of the processor according to a second embodiment of the present invention;

FIG. 5 is a block diagram showing the configuration of the processor according to a third embodiment of the present invention;

FIG. 6 is a block diagram showing the configuration of the processor according to a fourth embodiment of the present invention;

FIG. 7 is a block diagram showing the configuration of the processor according to a fifth embodiment of the present invention;

FIG. 8 is a block diagram showing the configuration of the processor according to a sixth embodiment of the present invention;

FIG. 9 is a block diagram showing the configuration of the processor according to a seventh embodiment of the present invention;

FIGS. 10A and 10B are diagrams showing an example of a fault relieving process when a fault has occurred, in the present invention; and

FIGS. 11A and 11B are diagrams showing another example of the fault relieving process when the fault has occurred, in the present invention.

Hereinafter, a processor of the present invention will be described in detail with reference to the attached drawings. The processor of the present invention is attained as a semiconductor integrated circuit in one chip. Also, the present invention will be described by using as an example, a case where the processor is used in a control system. However, the present invention is not limited to such an application.

FIG. 1 is a block diagram showing the configuration of the processor according to the first embodiment of the present invention. With reference to FIG. 1 , the processor in the first embodiment includes a memory 100 , a plurality of processing sections 102 A and 102 B, a plurality of fault detecting circuits 122 A and 122 B, a fault monitoring and control section 108 , and a selector section 106 .

The memory 100 is shared by the plurality of processing sections 102 A and 102 B and has memory regions 100 A and 100 B which are respectively used for the plurality of processing sections 102 A and 102 B. Also, the memory 100 has a memory region 100 Z to be commonly accessed by the plurality of processing sections 102 A and 102 B. Each of the memory regions 100 A, 100 B and 100 Z stores process programs A and B and Z, respectively, and also stores the data to be used when the programs are executed.

The processing section 102 A accesses the memory region 100 A of the memory 100 and executes the process program A stored therein. Thus, a desired process is executed. The process includes at least one processing unit. The processing section 102 A outputs a data related to the processing unit to the fault monitoring and control circuit 108 each time or the processing unit is completed. Also, the processing section 102 A outputs a selection instruction to the fault monitoring and control circuit 108 so that an input/output terminal Td is connected to the processing section 102 A through the selector section 106 , when an external data from an external device is required for the process or when a process result is outputted to the external device. When a relieving process instruction is received from the fault monitoring and control section 108 , the process program Z stored in the memory region 100 Z is executed instead of the process program A stored in the memory region 100 A. Also, the process program A may be executed. The processing section 102 B operates similarly to the processing section 102 A.

The fault detecting circuits 122 A and 122 B are provided for the processing sections 102 A and 102 B, respectively. Each of the fault detecting circuits 122 A and 122 B has a watch dog timer (not shown). The fault detecting circuit 122 A monitors and detects a fault of the processing section 102 A by using the watch dog timer and outputs an error detection signal to the fault monitoring and control circuit 108 . Similarly, the fault detecting circuit 122 B monitors and detects a fault of the processing section 102 B and outputs the error detection signal to the fault monitoring and control circuit 108 .

The selector section 106 is connected to a single input/output terminal TD, and connects the input/output terminal Td to one of the processing section 102 A and processing section 102 B in response to a selection control signal from the fault monitoring and control circuit 108 .

The fault monitoring and control circuit 108 generates the selection control signal in accordance with the selection instruction from each of the processing sections 102 A and 102 B and outputs to the selector section 106 . Thus, the input/output terminal Td is connected through the selector section 106 to the processing section 102 A or 102 B. Also, the fault monitoring and control circuit 108 detects the fault of any of the processing sections 102 A and 102 B in accordance with the error detection signal from the fault detecting circuit 122 A or 122 B and generates the selection control signal so that the input/output terminal Td is forced to be connected to the normal processing section. Also, the fault monitoring and control circuit 108 outputs the relieving process instruction to the normal processing section in response to the error detection signal. Also, for each completion of a predetermined processing unit, the fault monitoring and control circuit 108 receives the data related to the processing unit from each of the processing sections 102 A and 102 B. The fault monitoring and control circuit 108 can know what of the processing units is completed in each processor, based on this data.

The operation of the processor according to the first embodiment of the present invention will be described below with reference to FIG. 2 . At first, the processing sections 102 A and 102 B are activated to start executing the process programs A and B stored in the memory regions 100 A and 100 B, respectively. It is supposed that the process of the process program A and the process of the process program B are different. For example, the process of the process program A is a control of an external device (not shown) and the process of the process program B is a calculating process. Also, it is supposed that the processing sections 102 A and 102 B progresses execution of the process programs A and B in a predetermined process cycle. Moreover, it is supposed that the process program A includes a plurality of processing units A1, . . . , A2, . . . , and the process program B includes a plurality of processing units B1, . . . , B2, . . . . Here, in the processing unit A1, the process result is assumed to be outputted through the input/output terminal Td.

The processing section 102 A executes the process program A and completes the processing unit A1, and then outputs a selection instruction to the fault monitoring and control circuit 108 . The fault monitoring and control circuit 108 outputs the selection control signal to the selector section 106 in response to the selection instruction. As a result, the input/output terminal Td is connected through the selector section 106 to the processing section 102 A. The processing section 102 A outputs the process result of the processing unit A1 through the selector section 106 and the input/output terminal Td to the external device. At the same time, the processing section 102 A outputs the data related to the processing unit A1 to the fault monitoring and control circuit 108 . Thus, the fault monitoring and control circuit 108 can know the completion of the processing unit A1. The processing section 102 B similarly executes the processing unit B1.

Thereafter, it is supposed that the process in a process cycle T0 is completed and the control advances to the process in a process cycle T1. At this time, the processing section 102 A starts the process of the processing unit A2. Also, the processing section 102 B starts the process of the processing unit B2. At this time, when detecting an error of the processing section 102 A, the fault detecting circuit 122 A outputs the error detection signal to the fault monitoring and control circuit 108 . The fault monitoring and control circuit 108 outputs the relieving process instruction to the processing section 102 B in response to the error detection signal. In addition, the fault monitoring and control circuit 108 sends a progress state data, namely, the data related to the processing unit A2 in this example to the processing section 102 B to indicate what of the processing units has been completed. The processing section 102 B interrupts the processing unit B2 in the process program B and executes a relieving process program Z and then starts a relieving process.

FIG. 10A shows processing capacities of a CPU 1 and a CPU 2 included in the processing sections 102 A and 102 B. In a normal case, when the times assigned to the CPUs 1 and CPU 2 are T, the CPU 1 and CPU 2 execute the process programs A and B within a time of T/2 at the processing capacities of 100%. At this time, if a fault has occurred in the process of the CPU 2 and the process of the CPU 2 must be executed by the CPU 1 , the CPU 1 can execute both of the process programs A and B within the time T, as shown in FIG. 10B .

On the other hand, FIG. 11A shows the processing capacities of the CPU 1 and the CPU 2 included in the processing sections 102 A and 102 B. As shown in FIG. 11A , in a normal case, when the times assigned to the CPU 1 and CPU 2 are T, the CPU 1 and CPU 2 execute the process programs A and B within a time longer than the time T/2, and a time of T/2 in the processing capacities of 100%. At this time, if a fault has occurred in the CPU 2 and the process of the CPU 2 must be executed by the CPU 1 , the CPU 1 cannot execute both of the process programs A and B within the time of T, as shown in FIG. 11B . In such a case, some countermeasures are required such as a method in which the CPU 1 executes only the process program A or only the process program B or other methods.

In the relieving process, the importance/emergency of the processing unit A1 is determined. When the processing unit A1 is executed and the control is advanced, there is a case that a control system should be immediately stopped because of fault occurrence. Also, there is also a case that the subsequent processing unit A2 must be executed. In such a case, the entire control system cannot be stopped because of the fault in one processor. Next, the processing section 102 B determines a processing load of the processing section 102 A and a processing load of the processing section 102 B. For example, if the processing load of the processing section 102 A before the fault occurrence is 70% and the processing load of the processing section 102 B is 30%, the processing section 102 B can execute both of the process of the program A and the process of the program B. However, if the processing load of the processing section 102 A before the fault occurrence is 80% and the processing load of the processing section 102 B is 50%, the processing section 102 B cannot execute both of the processes of the programs A and B. In such a case, which of the processes should be executed is determined in accordance with the importance/emergency of the process of the program A and the process of the program B (the processing unit A2, . . . and the processing unit B2, . . . ).

FIG. 3 is a flowchart showing the reliving process based on the process program Z. With reference to FIG. 3 , at a step S 2 , the processing section 102 B acquires the loads of the processing sections 102 A and 102 B to be expected after the fault occurrence. The load of the processing section 102 A can be determined from the progress state data sent from the fault monitoring and control circuit 108 to indicate what processing unit has been completed, and the processing unit described in the process program A. Also, the load of the processing section 102 B can be determined from the processing unit of the process program B in accordance with the process state at the time of the fault occurrence. At this time, if the load of each processing unit is calculated in advance and held in a table (not shown) in the memory region 100 Z, it is possible to reduce the time required to calculate the load.

Next, the processing section 102 B determines the importance/emergency for each processing unit of the process programs A and B. For this purpose, it is adequate to prepare the table (not shown) in the memory region 100 Z in which a priority level is assigned in advance to each processing unit for the process programs A and B. This is because the processing unit contains an accessory processing unit. Since such a processing unit is not required to reserve the safety of the control system, its priority level is low. On the other hand, there is a case where, although the control process is advanced to the middle, the control is instable in that state and it must be advanced to a predetermined stable state, or a case that it must be returned to the stable state. For this reason, for example, the priority level of the process required to reserve the safety becomes high.

From the above reasons, at a step S 6 , whether or not the importance/emergency (priority) of the processing unit A2 is higher than that of the processing unit B2 is determined. If the importance/emergency (priority) of the processing unit A2 is higher, a step S 8 is executed. Then, whether or not the load becomes over 100% is determined when the processing unit A2 is executed by the processing section 102 B in addition to the processing unit to be originally executed. If the load is not over 100%, the processing unit A2 is executed at a step S 14 , and then the processing unit B2 is executed. When the processing unit A2 has been completed, the processing section 102 B outputs the process result of the processing unit A2 through the selector section 106 and the input/output terminal Td to the external device. At the same time, the processing section 102 B outputs a data related to the processing unit A2 to the fault monitoring and control circuit 108 . Thus, the fault monitoring and control circuit 108 can know the completion of the processing unit A2. At this time, by the fault monitoring and control circuit 108 , the input/output terminal Td is already connected through the selector section 106 to the processing section 102 B. Therefore, the processing section 102 B is not required to output the selection instruction to the fault monitoring and control circuit 108 .

Thereafter, whether or not the continuation of the process is possible is determined at a step S 16 . If the continuation is possible, the process is continued at a step S 18 . Also, if the continuation of the process is impossible, a step S 12 is executed to stop the control process.

Also, if the load is determined to be over 100%, the processing unit A2 is executed at a step S 10 . Thus, the safety is reserved. After that, the step S 12 is executed to stop the control process.

Also, if the importance/emergency of the processing unit A2 is low, whether or not the load is over 100% is determined at a step S 20 when the processing section 102 B executes the processing unit A2 in addition to the processing unit B2 to be originally executed. If the load is over 100%, the processing unit B2 is executed at a step S 22 . After that, the step S 12 is executed to stop the control process.

Also, if the load is determined not to be over 100% at the step S 20 , a step S 24 is executed. At the step S 24 , the processing unit B2 is executed, and then the processing unit A2 is executed. When the process of the processing unit A2 has been completed, the processing section 102 B outputs the process result of the processing unit A2 through the selector section 106 and the input/output terminal Td to the external device. At the same time, the processing section 102 B outputs the data related to the processing unit A2 to the fault monitoring and control circuit 108 . Thus, the fault monitoring and control circuit 108 can know the completion of the processing unit A2. At this time, by the fault monitoring and control circuit 108 , the input/output terminal Td is already connected through the selector section 106 to the processing section 102 B. Therefore, the processing section 102 B is not required to output the selection instruction to the fault monitoring and control circuit 108 .

Next, whether or not the continuation of the process is possible is determined at a step S 26 . If the continuation is possible, the process is continued at a step S 28 . Also, if the continuation of the process is impossible, a step S 22 is executed to stop the control process.

In this way, the processing section 102 B removes some of the processing units of the process programs A and B and can consequently advance the processes of the process programs A and B by using its usable processing performance.

In the above-mentioned example, since the process period is defined, the processing units are required to be completed by the processing sections 102 A and 102 B within the process period. However, if the process period is not defined, the processing section 102 B may execute the process programs A and B although the process time is necessary.

Also, in the above-mentioned example, the processing sections 102 A and 102 B output the selection instruction to the fault monitoring and control circuit 108 to control the selector section 106 . However, when the input/output terminal Td is always connected to the processing section 102 A, if it is only required to be connected to the processing section 102 B at the time of an error, the processing sections 102 A and 102 B are not required to output the selection instruction to the fault monitoring and control circuit 108 .

The processing sections 102 A and 102 B send the data related to the processing unit to the fault monitoring and control circuit 108 for each completion of the processing unit. However, for example, the processing section 102 A may directly send the data related to the processing unit to the processing section 102 B. Or, thee data may be written to the table (not shown) in the memory region 100 Z. In any of the cases, the processing sections 102 A and 102 B are not required to send the data related to the processing unit to the fault monitoring and control circuit 108 , for each completion of the processing unit.

Also, in the above-mentioned example, the single input/output terminal Td is connected through the selector section 106 to the processing sections 102 A and 102 B. However, a plurality of input/output terminals (not shown) may be designed to be connected to the processing sections 102 A and 102 B, respectively. In that case, the selector section 106 is not required, and the processes of the fault monitoring and control circuit 108 in association therewith are not required.

A processor according to the second embodiment of the present invention will be described below with reference to FIG. 4 . Attention should be paid to the fact that the memory 100 is not shown in FIG. 4 . Also, in the processor according to the second embodiment, the number of the processors is 2.

The processor in the second embodiment includes CPU 2 A and CPU 2 B, error detecting circuits 22 A and 22 B, buses 12 A and 12 B, a bus bridge circuit 10 , input/output ports 4 - 1 -A, 4 - 2 -A, 4 - 1 -B and 4 - 2 -B, error detecting circuits 24 A and 24 B, a selector section 6 and an fault monitoring and control circuit 8 a . Here, the fault monitoring and control circuit 8 a corresponds to the fault monitoring and control circuit 108 in the first embodiment, and the selector section 6 corresponds to the selector section 106 in the first embodiment. In the following embodiment, although the fault monitoring and control circuits are referred to as 8 a , 8 b , 8 c , 8 d , 8 e and 8 f , they are similar. The error detecting circuits 22 A and 24 A correspond to the fault detecting circuit 122 A in the first embodiment, and the error detecting circuits 22 B and 24 B correspond to the fault detecting circuit 122 B in the first embodiment. Also, the CPU 2 A, the bus 12 A, the bus bridge circuit 10 and the input/output ports 4 - 1 -A, 4 - 2 -A correspond to the processing section 102 A in the first embodiment, and the CPU 2 B, the bus 12 B, the bus bridge circuit 10 and the input/output ports 4 - 1 -B, 4 - 2 -B correspond to the processing section 102 B in the first embodiment. Also, FIG. 4 does not show a signal line for the relieving process instruction from the fault monitoring and control circuit 8 a to the CPU 2 A and CPU 2 B; a signal line for the selection instruction from the CPU 2 A and CPU 2 B to the fault monitoring and control circuit 8 a ; and a signal line for the data related to the processing unit when the processing unit has been completed.

The CPU 2 A and CPU 2 B execute the process programs A and B stored in the memory (not shown), respectively. The error detecting circuits 22 A and 22 B are provided for the CPU 2 A and CPU 2 B, respectively. Each of the error detecting circuits 22 A and 22 B has a watch dog timer and monitors the error of the corresponding CPU. When the error is detected, the error detection signal is outputted to the fault monitoring and control circuit 8 a.

The error detecting circuit 24 A is provided for the input/output ports 4 - 1 -A and 4 - 2 -A, and the error detecting circuit 24 B is provided for the input/output ports 4 - 1 -B and 4 - 2 -B. Each of the error detecting circuits 24 A and 24 B determines that the state in which a fixed value is always outputted from each input/output port is an error (fault) state, and then outputs the error detection signal to the fault monitoring and control circuit 8 a.

The input/output ports 4 - 1 -A and 4 - 2 -A are connected to the bus 12 A, and the input/output ports 4 - 1 -B and 4 - 2 -B are connected to the bus 12 B. The bus bridge circuit 10 connects the buses 12 A and 12 B and is connected to the CPU 2 A and CPU 2 B. Thus, the input/output ports 4 - 1 -A and 4 - 2 -A and the input/output ports 4 - 1 -B and 4 - 2 -B are connected to the CPU 2 A and CPU 2 B.

The selector section 6 has a buffer driver 62 , an AND gate 72 , an OR gate 78 , selectors 64 , 68 , 74 and 76 . Input enable signals ENI of the input/output ports 4 - 1 -A and 4 - 1 -B are connected to an input of the OR gate 78 and an input of the selector 76 . An output of the selector 76 and an output of the OR gate 78 are connected to inputs of the selector 74 . An output of the selector 74 is connected to one input of the AND gate 72 . An input/output terminal Td is connected to the other input of the AND gate 72 . An output of the AND gate 72 is connected to data inputs DIN of the input/output ports 4 - 1 -A and 4 - 1 -B. Output enable signals ENO of the input/output ports 4 - 1 -A and 4 - 1 -B are connected to inputs of the selector section 64 . An output of the selector section 64 is connected to a control terminal of the buffer driver 62 . Data outputs DOUT of the input/output ports 4 - 1 -A and 4 - 1 -B are connected to inputs of the selector section 68 . An output of the selector section 68 is connected to an input of the buffer driver 62 , and an output of the buffer driver 62 is connected to the input/output terminal Td. A selection control signal is sent from the fault monitoring and control circuit 8 a to each of the selectors 64 , 68 , 74 and 76 . As mentioned above, the input/output terminal Td is connected to the input/output port 4 - 1 -A in the normal state and connected to the input/output port 4 - 1 -B in the fault state.

The fault monitoring and control circuit 8 a has a detection signal decoder 82 a and a switching signal output circuit 84 a . The detection signal decoder 82 a decodes the error detection signals from the error detecting circuits 22 A and 22 B, 24 A and 24 B, generates a fault decode resultant signal and outputs to the switching signal output circuit 84 a . The switching signal output circuit 84 a outputs the selection control signal to the selector section 6 in response to the fault decode resultant signal.

The operation of the processor according to the second embodiment is similar to that of the processor in the first embodiment. Thus, their explanations are omitted.

A processor according to the third embodiment of the present invention will be described below with reference to FIG. 5 . The processor in the third embodiment is similar to the processor according to the second embodiment. The processor in the third embodiment is different from the processor in the second embodiment in that instead of the buses 12 A and 12 B and the bus bridge circuit 10 , only the buses 12 A and 12 B are used and error detecting circuits 10 A and 10 B are provided for them, respectively. When a bus error has occurred, the data cannot be received by and outputted from the CPU. For this reason, the error detecting circuits 10 A and 10 B are provided. The error detecting circuits 10 A and 10 B, for example, if the data on the bus is always at “1” or “0”, determines the bus as the error and outputs the error detection signal to the fault monitoring and control circuit 8 b . The fault monitoring and control circuit 8 b is similar in configuration and operation to the fault monitoring and control circuit 8 a.

The operation of the processor according to the third embodiment is similar to that of the processor in the first embodiment. Thus, their explanations are omitted.

A processor according to a fourth embodiment of the present invention will be described below with reference to FIG. 6 . The processor according to the fourth embodiment is similar to the processor according to the second embodiment. The processor according to the fourth embodiment is different from the processor according to the second embodiment in configuration and operation of a fault monitoring and control circuit 8 c . In the fourth embodiment, the fault monitoring and control circuit 8 c has a detection signal decoder 82 c , a switching signal output circuit 84 c and an internal memory 86 c . The detection signal decoder 82 c decodes the error detection signals from the error detecting circuits 22 A and 22 B, 24 A and 24 B, generates the fault decode resultant signal and outputs to the switching signal output circuit 84 a . The switching signal output circuit 84 a outputs the selection control signal to the selector section 6 in response to the fault decode resultant signal. The internal memory 86 c stores the fault decode resultant signal. Consequently, when the fault will be later analyzed, it is possible to quickly determine the occurrence position of the error by examining the data stored in the internal memory 86 c . Also, the fault decode resultant signal is outputted from an output terminal Te to the external device. The external device can know the occurrence of the error in the processor by monitoring and detecting the signal of the output terminal Te.

The operation of the processor according to the fourth embodiment is similar to that of the first embodiment. Thus, their explanations are omitted.

The processor according to the fifth embodiment of the present invention will be described below with reference to FIG. 7 . The processor according to the fifth embodiment is similar to the processor according to the third embodiment. The processor according to the fifth embodiment is different from the processor according to the third embodiment in configuration and operation of a fault monitoring and control circuit 8 d . In the fifth embodiment, the fault monitoring and control circuit 8 d has a detection signal decoder 82 d , a switching signal output circuit 84 d , an internal memory 86 d and an access control circuit 88 d . The detection signal decoder 82 d decodes the error detection signals from the error detecting circuits 22 A and 22 B, 24 A and 24 B, 10 A and 10 B, generates the fault decode resultant signal and outputs to the switching signal output circuit 84 d , the internal memory 86 d and the access control circuit 88 d . The switching signal output circuit 84 d outputs the selection control signal to the selector section 6 in response to the fault decode resultant signal. The internal memory 86 d stores the fault decode resultant signal. Consequently, when the error will be later analyzed, it is possible to quickly determine the occurrence position of the error by examining the data stored in the internal memory 86 d . Also, the fault decode resultant signal is outputted to the access control circuit 88 d . The access control circuit 88 d is connected to the buses 12 A and 12 B. The CPU 2 A and CPU 2 B can access the access control circuit 88 d through the buses 12 A and 12 B. The access control circuit 88 d reads the data from the internal memory 86 d in accordance with a command from the CPU 2 A and CPU 2 B, and outputs to a command issuing source. However, the access control circuit 88 d rejects the access from the CPU on the error occurrence side in accordance with the fault decode resultant signal from the detection signal decoder 82 d . When the CPU on the normal side executes the relieving process in response to the relieving process instruction, the necessary data can be extracted from the internal memory 86 d.

The other operation of the processor according to the fifth embodiment is similar to that of the processor in the first embodiment. Thus, their explanations are omitted.

The processor according to the sixth embodiment of the present invention will be described below with reference to FIG. 8 . The processor according to the sixth embodiment is similar to the processor according to the fifth embodiment. The processor according to the sixth embodiment is different from the processor according to the fifth embodiment in that RESET registers 16 A and 16 B are provided to be connected to the buses 12 A and 12 B. The CPU 2 A sets a reset command in the RESET register 16 A through the bus 12 A, and the CPU 2 B is reset by this reset command. Also, the CPU 2 B sets the reset command in the RESET register 16 B through the bus 12 B, and the CPU 2 A is reset by this reset command. In this way, for example, when the CPU 2 A is frozen, the CPU 2 B responds to the relieving process instruction from a fault monitoring and control circuit 8 e and sets the reset command in the register 16 B and then resets the CPU 2 A. Consequently, the operation of the CPU 2 A is expected to be returned, and the CPU 2 A may be returned to the normal operation.

The other operation of the processor according to the sixth embodiment is similar to that of the processor in the first embodiment. Thus, their explanations are omitted.

The processor according to the seventh embodiment of the present invention will be described below with reference to FIG. 9 . The processor according to the seventh embodiment is similar to the processor according to the sixth embodiment. The processor according to the seventh embodiment is different from the processor according to the sixth embodiment in the following manner. That is, an oscillator 15 A generates a clock signal with a high frequency and sends to the CPU 2 A. The CPU 2 A operates in synchronization with the clock signal with the high frequency. An oscillation monitoring circuit 26 A is provided for this oscillator 15 A, and when detecting an oscillation error, outputs the error detection signal to a fault monitoring and control circuit 8 f . Also, an oscillator 15 B generates a clock signal with a low frequency and sends to the CPU 2 B. The CPU 2 B operates in synchronization with the clock signal of the low frequency. An oscillation monitoring circuit 26 B is provided for this oscillator 15 B and when detecting the oscillation error, outputs the error detection signal to the fault monitoring and control circuit 8 f . Also, a low voltage detecting circuit 9 f is provided for a power source (not shown). Then, when a power source voltage becomes lower than a predetermined level, the error detection signal is outputted to the fault monitoring and control circuit 8 f . When the error detection signal is received from the low voltage detecting circuit 9 f , the fault monitoring and control circuit 8 f outputs the relieving process instruction to the CPU 2 B. Instead of the CPU 2 A, the CPU 2 B executes the process program A. In this way, before the power source voltage becomes lower than an operable voltage, the necessary process can be executed.

The other operation of the processor according to the sixth embodiment is similar to that of the processor in the first embodiment. Thus, their explanations are omitted.

As mentioned above, according to the present invention, depending on the control target, even if one of the plurality of CPUs is inoperable, the minimum process can be executed until the safety is reserved. Also, even if the processing performance is reduced, the main process can continue to be executed.

Also, when the CPU issues a request signal to an arbitration circuit of the bus, the arbitration circuit executes a bus arbitration and gives a bus use right (grant signal) to the CPU. When the bus is opened, the CPU issues a completion signal to the arbitration circuit. Then, the bus is opened in accordance therewith. When the CPU reserves the use right of the bus, even if an error is caused not to allow the bus arbitration or the CPU switching, the process is executed by the CPU other than the CPU in which the error has occurred, or since the reset is performed, the down of the entire system is prevented in the present invention.

Also, in a semiconductor integrated circuit, the reduction in the number of the terminals is desired. In that case, conventionally, when a particular terminal is fixedly assigned to a particular CPU, and the error has occurred in the CPU, the other CPU cannot be connected to the terminal even if the other CPU tries to continue the process. Thus, the data cannot be obtained. However, according to the present invention, at the time of the error occurrence, the substitution CPU can obtain the data through the terminal and continue the process. In this way, the flexibility in use of the processor can be increased.

According to the present invention, in the multi-processor system, even if an error has occurred in any CPU, it is possible to avoid immediate system down. Also, it is possible to execute a minimum process to reserve safety. It is possible to continue a necessary process based on a load state. Also, even if the number of the terminals is decreased, the terminals are shared, which allows the signal to be transmitted to and received from an external device by the normal CPU, even when the error has occurred. Also, it is possible to design a flexible system configuration through the sharing.